Deep real-world drug discovery — our investment in Character Biosciences
By Joel Dudley and Nick Olsen
We have only just begun to realize the value of human genome-enabled therapeutic discovery. Scientists completed the first draft of the human genome two decades ago (and just recently completed it). The initial sequencing of the human genome catalyzed great optimism for developing and tailoring treatments for each patient based on their unique biology. Genome scientists quickly realized that genomic sequence information was only part of a more complicated puzzle that comprises the vision for precision medicine. Precision oncology is likely our best example of precision medicine in the clinic. Tumor DNA sequencing is becoming more routine in clinical care and transforming how we treat patients and develop new medicines. However, even in oncology DNA sequencing alone often fails to capture a complete picture of a patient’s disease where causal genetic changes typically drive tumorigenesis. We find excitement in efforts to build multi-omic pictures of a patient’s disease, including RNA sequencing, spatial biology, proteomics, and clinical information on co-morbidities, among many other relevant attributes. Companies like Tempus and Roche (which acquired both Flatiron and Foundation Medicine) are pursuing this path.
Drug discovery efforts to explore human genetics to identify therapeutic targets from human populations continue to gain traction. The discovery of PCSK9 as a target for lipid-lowering therapies is perhaps one of the more famous examples of human genetics-guided drug discovery. Despite success stories and early wins, there remains plenty of room for improvement and innovation. Researchers use genome-wide association studies (GWAS) to identify genomic regions and specific genetic variants associated with human diseases. The results of genomic association studies are generally only as valid as the phenotypes defined in the study design. My academic GWAS studies define fairly simplistic, reductionist phenotypes that often only describe disease status (e.g., individuals exhibiting a particular disease phenotype vs. controls). However, disease phenotypes in the “real world” are usually more complex and nuanced in their description and manifestation. Therefore, we need well-characterized phenotypes to build richer genotype-to-phenotype maps to investigate the genetic basis of human disease. We need significantly more data about the unique complexity of each phenotype that is not available to most researchers. For example, we need higher-dimensional representations of disease, such as clinical annotations that go beyond the ICD-10 codes used for diagnosis and billing. We need longitudinal representations of the trajectory of diseases and their myriad manifestations over time. We need multi-modal data sets that incorporate clinical imaging and other high-dimensional representations of disease state in specific organs and tissues. We must embrace rather than reduce the complexity of human disease biology in real-world patient populations to realize the full potential of human genetics in therapeutic discovery.
This is the essence of therapeutic discovery based on deep phenotyping. And this is the mission of Character Biosciences.
We believe that the next frontier of precision medicine lies at the convergence of genomics, deep phenotyping, and machine learning applied to complex polygenic disorders (a genetic disorder caused by more than one gene). The way forward leverages advanced technologies to subtype patients in clinically meaningful populations and then treats each patient in the context of the disease they uniquely present. A fundamentally different organization and strategy are required to effectively enable this vision — one that is integrated across our siloed healthcare ecosystem; that does the hard work of acquiring, cleaning, transforming, and merging large, distributed data assets; that works directly with patients, providers, payers, and everyone in between; and that fundamentally thinks about discovering new treatments differently than the historical paradigms of drug discovery. And you need a team that can bring together these historically disparate capabilities — from diagnostics and drug discovery, to statistical genetics and computational biology, to patient advocacy and provider partnerships, to data integration and machine learning.
Character Biosciences has been assembling a world-class team and building an integrated precision medicine platform focused on treating age-related diseases with the highest unmet needs. They are building large integrated clinical genomic databases through partnerships with patients, healthcare providers, payers, and other industry partners. Their platform for drug and diagnostic discovery combines genetics, clinical and imaging data, computational biology, machine learning, and novel experimental approaches to disentangle complex genetically-associated diseases of aging.
The end goals are deeply phenotyped patient populations, treatments matched to their unique disease, companion diagnostics to identify and treat in the clinic, and direct relationships with providers and patients to speed the time to diagnosis and clinical impact.
We are proud to partner with Cheng, Marcel, Erik, Nick, and the rest of the Character Biosciences team. It’s a long journey ahead to deliver meaningful innovations to patients, but one we’re committed to embarking on together. We’re excited to work alongside our partners at Catalio Capital Management, Section 32, Casdin Capital, Lifeforce Capital, KdT Ventures, Cantos Ventures, Industry Ventures, and Clover Health.
If you’re a patient or know someone with age-related macular degeneration (AMD), help contribute to research here. If you’re a provider and determined to discover new treatments, please reach out here.
–
